Perpenidcularity

[[Am...]]

Alright guys I am looking for some opinions here, i have a program that we have been running for at least two years now and this has never been an issue but now that we have a batch thats failing a perpendicularity call out everyone has an opinion on how to measure this and i just feel like everyone is jumping through hoops to get passing parts because they don't want to 100%.

It is a molded part with a machined flat surface, that surface needs to be perpendicular within .002" to an Id right through the part. The diameter is about .526" and I have 3 circles in the cylinder evenly spaced. We have always average around .0009" on this dimension now getting about .0019"-.0021". The ID itself has a ± .001". The surface also has a .001" flatness call out that is around .0004-.0005.

What Im being asked to do now is to change the program and I just hate that we are changing the program to a single batch of parts when we have been using it this way for a while probably going back 20,000 parts. Ultimately the call is going to be out of my hands but I want to get some opinions on what are the objective wrong ways to do this.

One of the suggestions is to use the Maximum Inscribed element evaluation for the cylinder, another is checking a 2D line down the side of the ID, I don't like this as I know the result is going to vary based on where I put that line and Ive evaluated the shape of this ID years back, after the shrink from molding and a 21 day oven cycle to cure the ID does have aslght bow in the very middle being slightly larger then the rest. One thing Ive noticed about these parts is that the ID itself is actually slightly smaller then its been over the last two years and this material tends to wear out core pins so for me this is within reason that the core pin has been worn out and is not as perpendicular as it used to be.

What would you guy do and what is the correct way to check this perpendicularity? I am going to have to play with this at some point today to please production and I just want to hear some opinions and what methods to avoid entirely.

[[Ma...]]

If calibration of pins don't work, then I presume that something has changed in technology of part or machine wear.
Two years of same program can not change results by its own. Better to check on another machine or use reference part.

[[Am...]]

Same results on my duramax and on O-inspect, same after qualification. The issue is that to avoid 100% they are evaluating if we are measuring this perpendicularity correctly.

So im wondering what others would do in thier construction of this perpendicularity, what features to use and what evaluation methods.

[[Da...]]

It's been a couple year since I played with Perp. I'm using Calypso 6.2 so it may calculate different now. I'm thinking Calypso creates square plane the length and width of your points splay. Then when it calculates perp it basically gives the plane corner to corner distance as perp. If your plane is perfectly square but has a dip or crown, that flatness error won't be reported as TIR it would give a zero perp. I use it some with discretion. Most of the time I take individual points (don't print all the points) Then I do a minimum and maximum feature, then do a distance between the those 2 points. That gives you the TIR that is an accurate result. Using the perp in Calypso can give you a result that is too hi or too low. On the bore pickup... I tend to do my Cylinder circle paths at the top and bottom of the cylinder and none in middle. Not for size but for the cylinder axis that I'm going to reference from. Maybe on molded or not real precisely formed bores, take 2 circle paths at top and 2 at bottom maybe .03" apart. You definitely could get many opinions on the bore pickup and all could be pretty good.

[[Am...]]

Is it appropriate to be using Outer Tangential evaluation on the Datum Cylinder here?

[[Da...]]

I wouldn't use Outer Tangential....Even the zeiss manual says its not as stable or repeatable (probably not in those words).. Not really good for ID., in fact I never use it myself. Your max inscribed simulates a pin better. LSQ may be the most repeatable.

[[Ma...]]

Amel,

Is it possible to show us the section of the drawing or at least provide a sketch with pertinent dimensions and GDT?

Also, you don't say how you are currently evaluating the bore. Are you using a cylinder or a 3d Line?

For a bore as you describe, Outer Tangential Cylinder and Maximum Inscribed Cylinder are the same. This may be the technically correct method for a datum but not the most repeatable.

[[Am...]]

I am currently using a Cylinder with 3 circle paths, I have done it with Maximum Inscribed evaluation and one with LSQ both have the same repeatability but the LSQ evaluation is coming up right at the top of the tolerance and sometimes over, the max inscribed is averaging about .001" and the LSQ about .0019" with some parts being .0022".

Datum A is the bore, the plane is machined to a flatness of about .0004" allowed to be .001" and needs to be perpendicular to A within .002"

Datum A is not machined it is molded so it has a slightly wider middle then the top and bottom but I imagine using the Maximum Inscribed circle that should not make much of a difference and I think thats why the LSQ is most likely highe because the form of the molded cylinder is affecting the axis of the cylinder.

Would it be better to use a 3D line here? should I avoid the middle entirely? Because its wider I think if the application has a pin going through there I don't think that middle circle will matter in the end but I think is effecting the axis for this perpendicularity.

I will try to get a sketch of the relevant features and the callout, I cannot share the print.